How to measure time taken by a function to execute

I need to get execution time in milliseconds.

I originally asked this question back in 2008. The accepted answer then was to use new Date().getTime() However, we can all agree now that using the standard API is more appropriate. I am therefore changing the accepted answer to this one.


var t0 =

doSomething()   // <---- measured code goes between t0 and t1

var t1 =
console.log("Call to doSomething took " + (t1 - t0) + " milliseconds.")

In Node.js it is required to import the performance class

Using console.time: (living standard)


doSomething()   // <---- The function you're measuring time for 


The string being passed to the time() and timeEnd() methods must match
(for the timer to finish as expected).

console.time() documentations:

use new Date().getTime()

The getTime() method returns the number of milliseconds since midnight of January 1, 1970.


var start = new Date().getTime();

for (i = 0; i < 50000; ++i) {
// do something

var end = new Date().getTime();
var time = end - start;
alert('Execution time: ' + time);

Don’t use Date(). Read below.


var a =;
alert('do something...');
var b =;
alert('It took ' + (b - a) + ' ms.');

It works on:

  • IE 10 ++

  • FireFox 15 ++

  • Chrome 24 ++

  • Safari 8 ++

  • Opera 15 ++

  • Android 4.4 ++

  • etc, etc

console.time may be viable for you, but it’s non-standard §:

This feature is non-standard and is not on a standards track. Do not use it on production sites facing the Web: it will not work for every user. There may also be large incompatibilities between implementations and the behavior may change in the future.

Besides browser support, seems to have the potential to provide more accurate timings as it appears to be the bare-bones version of console.time.

<rant> Also, DON’T EVER use Date for anything because it’s affected by changes in “system time”. Which means we will get invalid results —like “negative timing”— when the user doesn’t have an accurate system time:

On Oct 2014, my system clock went haywire and guess what…. I opened Gmail and saw all of my day’s emails “sent 0 minutes ago“. And I’d thought Gmail is supposed to be built by world-class engineers from Google…….

(Set your system clock to one year ago and go to Gmail so we can all have a good laugh. Perhaps someday we will have a Hall of Shame for JS Date.)

Google Spreadsheet’s now() function also suffers from this problem.

The only time you’ll be using Date is when you want to show the user his system clock time. Not when you want to get the time or to measure anything.

If you need to get function execution time on your local development machine, you can either use your browser’s profiling tools, or console commands such as console.time() and console.timeEnd().

All modern browsers have JavaScript profilers built-in. These profilers should give the most accurate measurement as you do not have to modify your existing code, which could affect the function’s execution time.

To profile your JavaScript:

  • In Chrome, press F12 and select the Profiles tab, then Collect JavaScript CPU Profile.
  • In Firefox, install/open Firebug, and click on the Profile button.
  • In IE 9+, press F12, click on Script or Profiler (depending on your version of IE).

Alternatively, on your development machine, you can add instrumentation to your code with console.time() and console.timeEnd(). These functions, supported in Firefox11+, Chrome2+ and IE11+, report on timers that you start/stop via console.time(). time() takes a user-defined timer name as an argument, and timeEnd() then reports on the execution time since the timer started:

function a() {
  ... do stuff ...
  var dur = console.timeEnd("myTimer"); // NOTE: dur only works in FF

Note that only Firefox returns the elapsed time in the timeEnd() call. The other browsers simply report the result to the developer console: the return value of timeEnd() is undefined.

If you want to get function execution time in the wild, you will have to instrument your code. You have a couple options. You can simply save the start and end times by querying new Date().getTime():

function a() {
  var start = new Date().getTime();
  ... do stuff ...
  var end = new Date().getTime();
  var dur = end - start;

However, the Date object only has millisecond resolution and will be affected by any OS’s system clock changes. In modern browsers, there’s a better option.

The better option is to use the High Resolution Time, aka now() is better than the traditional Date.getTime() in two important ways:

  1. now() is a double with submillisecond resolution that represents the number of milliseconds since the start of the page’s navigation. It returns the number of microseconds in the fractional (e.g. a value of 1000.123 is 1 second and 123 microseconds).

  2. now() is monotonically increasing. This is important as Date.getTime() can possibly jump forward or even backward on subsequent calls. Notably, if the OS’s system time is updated (e.g. atomic clock synchronization), Date.getTime() is also updated. now() is guaranteed to always be monotonically increasing, so it is not affected by the OS’s system time — it will always be wall-clock time (assuming your wall clock is not atomic…).

now() can be used in almost every place that new Date().getTime(), + new Date andt are. The exception is that Date and now() times don’t mix, as Date is based on unix-epoch (the number of milliseconds since 1970), while now() is the number of milliseconds since your page navigation started (so it will be much smaller than Date).

Here’s an example of how to use now():

function a() {
  var start =;
   ... do stuff ...
  var end =;
  var dur = end - start;

now() is supported in Chrome stable, Firefox 15+, and IE10. There are also several polyfills available.

One other option for measuring execution time in the wild is UserTiming. UserTiming behaves similarly to console.time() and console.timeEnd(), but it utilizes the same High Resolution Timestamp that now() uses (so you get a sub-millisecond monotonically increasing clock), and saves the timestamps and durations to the PerformanceTimeline.

UserTiming has the concepts of marks (timestamps) and measures (durations). You can define as many of either as you want, and they’re exposed on the PerformanceTimeline.

To save a timestamp, you call mark(startMarkName). To get the duration since your first mark, you simply call measure(measurename, startMarkname). The duration is then saved in the PerformanceTimeline alongside your marks.

function a() {
  ... do stuff ...
  window.performance.measure("myfunctionduration", "start");

// duration is window.performance.getEntriesByName("myfunctionduration", "measure")[0];

UserTiming is available in IE10+ and Chrome25+. There is also a polyfill available (which I wrote).

To get precise values you should use Performance interface. It’s supported in modern versions of Firefox, Chrome, Opera and IE. Here’s an example of how it can be used:

var performance = window.performance;
var t0 =;
var t1 =;
console.log("Call to doWork took " + (t1 - t0) + " milliseconds.")

Date.getTime() or console.time() are not good for measuring precise execution time. You can use them if quick rough estimate is OK for you. By rough estimate I mean you can get 15-60 ms shift from the real time.

Check this brilliant post on measuring execution time in JavaScript. The author also gives a couple of links about accuracy of JavaScript time, worth reading.

Use Firebug, enable both Console and Javascript. Click Profile. Reload. Click Profile again. View the report.

A simple solution, you can use add operator also here

 var start = +new Date();
 var end = +new Date();
 var time = end - start;
 console.log('total execution time="+ time + "ms');

var StopWatch = function (performance) {
    this.startTime = 0;
    this.stopTime = 0;
    this.running = false;
    this.performance = performance === false ? false : !!window.performance;

StopWatch.prototype.currentTime = function () {
    return this.performance ? : new Date().getTime();

StopWatch.prototype.start = function () {
    this.startTime = this.currentTime();
    this.running = true;

StopWatch.prototype.stop = function () {
    this.stopTime = this.currentTime();
    this.running = false;

StopWatch.prototype.getElapsedMilliseconds = function () {
    if (this.running) {
        this.stopTime = this.currentTime();

    return this.stopTime - this.startTime;

StopWatch.prototype.getElapsedSeconds = function () {
    return this.getElapsedMilliseconds() / 1000;

StopWatch.prototype.printElapsed = function (name) {
    var currentName = name || 'Elapsed:';

    console.log(currentName, '[' + this.getElapsedMilliseconds() + 'ms]', '[' + this.getElapsedSeconds() + 's]');


var stopwatch = new StopWatch();

for (var index = 0; index < 100; index++) {
    stopwatch.printElapsed('Instance[' + index + ']');




Instance[0] [0ms] [0s]
Instance[1] [2.999999967869371ms] [0.002999999967869371s]
Instance[2] [2.999999967869371ms] [0.002999999967869371s]
/* ... */
Instance[99] [10.999999998603016ms] [0.010999999998603016s]
Elapsed: [10.999999998603016ms] [0.010999999998603016s] is optional – just pass false into StopWatch constructor function.

process.hrtime() is available within Node.js – it returns a value in nanoseconds

var hrTime = process.hrtime()
console.log(hrTime[0] * 1000000 + hrTime[1] / 1000)

It is possible to use only one variable:

var timer =;

// Do something

timer +=;
console.log("Time: " + (timer/1000).toFixed(5) + " sec.")

timer/1000 – to convert milliseconds to seconds

.toFixed(5) – to trim extra digits

To extend vsync’s code further to have the ability to return the timeEnd as a value in NodeJS use this little piece of code.

console.timeEndValue = function(label) { // Add console.timeEndValue, to add a return value
   var time = this._times[label];
   if (!time) {
     throw new Error('No such label: ' + label);
   var duration = - time;
   return duration;

Now use the code like so:

console.time('someFunction timer');


var executionTime = console.timeEndValue('someFunction timer');
console.log("The execution time is " + executionTime);

This gives you more possibilities. You can store the execution time to be used for more purposes like using it in equations, or stored in a database, sent to a remote client over websockets, served on a webpage, etc.

Here’s a decorator for timing functions

let timed = (f) => (...args)=>{
    let start =;
    let ret = f(...args);
    console.log(`function ${} took ${(}ms`)
    return ret;   


let test = ()=>{/*does something*/}
test = timed(test)   // turns the function into a timed function in one line
test()               // run your code as normal, logs 'function test took 1001.900ms' 

If you’re using async functions you can make timed async and add an await before f(…args), and that should work for those. It gets more complicated if you want one decorator to handle both sync and async functions.

there are multiple ways to achieve this objective :

  1. using console.time

    //run the function in between these two lines for that you need to 
    //measure time taken by the function. ("ex. function();")
  2. this is the most efficient way :
    using, e.g.

    var v1 =;
    //run the function here for which you have top measure the time 
    var v2 =;
    console.log("total time  taken = "+(v2-v1)+"milliseconds");
  3. use +(add operator) or getTime()

    var h2 = +new Date(); //or
    var h2 = new Date().getTime();
    for(i=0;i<500;i++) { /* do something */}
    var h3 = +new Date();   //or 
    var h3 = new Date().getTime();
    var timeTaken = h3-h2;
    console.log("time ====", timeTaken);

Here’s what happens when you apply the unary plus operator to a Date instance:
Get the value of the Date instance in question
Convert it to a Number

NOTE: getTime() gives better performance than unary + operator.

Since console.time and aren’t supported in some major browsers (i.e. IE10), I created a slim utility that utilizes the best available methods. However, it lacks error handling for false usages (calling End() on a not initialized timer).

Use it and improve it as you want.

Performance: {
    Timer: {},
    Start: function (name) {
        if (console && console.time) {
        } else if ( {
            this.Timer[name] =;
        } else {
            this.Timer[name] = new Date().getTime();
    End: function (name) {
        if (console && console.time) {
        } else {
            var result;
            if ( {
                result = - this.Timer[name];
            } else {
                result = new Date().getTime() - this.Timer[name];
            console.log(name + ": " + result);

It may help you.

var t0 =;
var t1 =;
console.log("Call to doSomething took approximate" + (t1 - t0)/1000 + " seconds.")

Thanks, Achim Koellner, will expand your answer a bit:

var t0 = process.hrtime();
//Start of code to measure

//End of code
var timeInMilliseconds = process.hrtime(t0)[1]/1000000; // dividing by 1000000 gives milliseconds from nanoseconds

Please, note, that you shouldn’t do anything apart from what you want to measure (for example, console.log will also take time to execute and will affect performance tests).

Note, that in order by measure asynchronous functions execution time, you should insert var timeInMilliseconds = process.hrtime(t0)[1]/1000000; inside the callback. For example,

var t0 = process.hrtime();
someAsyncFunction(function(err, results) {
var timeInMilliseconds = process.hrtime(t0)[1]/1000000;


A couple months ago I put together my own routine that times a function using — even though at the time the accepted method seemed to be —
because the performance object is not yet available (built-in) in the stable Node.js release.

Today I was doing some more research and found another method for timing. Since I also found how to use this in Node.js code, I thought I would share it here.

The following is combined from the examples given by w3c and Node.js:

function functionTimer() {
    performance.measure('Start to End', 'start', 'end')
    const measure = performance.getEntriesByName('Start to End')[0]


If you intend to use the performance object in a Node.js app, you must include the following require:

const { performance } = require('perf_hooks')

Stopwatch with cumulative cycles

Works with server and client (Node or DOM), uses the Performance API.
Good when you have many small cycles e.g. in a function called 1000 times that processes 1000 data objects but you want to see how each operation in this function adds up to the total.

So this one uses a module global (singleton) timer. Same as a class singleton pattern, just a bit simpler to use, but you need to put this in a separate e.g. stopwatch.js file.

const perf = typeof performance !== "undefined" ? performance : require('perf_hooks').performance;
const DIGITS = 2;

let _timers = {};

const _log = (label, delta?) => {
    if (_timers[label]) {
        console.log(`${label}: ` + (delta ? `${delta.toFixed(DIGITS)} ms last, ` : '') +
            `${_timers[label].total.toFixed(DIGITS)} ms total, ${_timers[label].cycles} cycles`);

export const Stopwatch = {
    start(label) {
        const now =;
        if (_timers[label]) {
            if (!_timers[label].started) {
                _timers[label].started = now;
        } else {
            _timers[label] = {
                started: now,
                total: 0,
                cycles: 0
    /** Returns total elapsed milliseconds, or null if stopwatch doesn't exist. */
    stop(label, log = false) {
        const now =;
        if (_timers[label]) {
            let delta;
            if(_timers[label].started) {
                delta = now - _timers[label].started;
                _timers[label].started = null;
                _timers[label].total += delta;
            log && _log(label, delta);
            return _timers[label].total;
        } else {
            return null;
    /** Logs total time */
    log: _log,
    delete(label) {
        delete _timers[label];

export default class Singleton {

  static myInstance: Singleton = null;

  _timers: any = {};

   * @returns {Singleton}
  static getInstance() {
    if (Singleton.myInstance == null) {
      Singleton.myInstance = new Singleton();

    return this.myInstance;

  initTime(label: string) {
    this._timers[label] =;
    return this._timers[label];

  endTime(label: string) {
    const endTime =;
    if (this._timers[label]) {
      const delta = endTime - this._timers[label];
      const finalTime = `${label}: ${delta}ms`;
      delete this._timers[label];
      return finalTime;
    } else {
      return null;

InitTime related to string.

return Singleton.getInstance().initTime(label); // Returns the time init

return Singleton.getInstance().endTime(label); // Returns the total time between init and end

If you want to measure the time between multiple things that aren’t nested you could use this:

function timer(lap){ 
    if(lap) console.log(`${lap} in: ${(}ms`); 
    timer.prev =;

Similar to console.time(), but easier usage if you don’t need to keep track of previous timers.

If you like the blue color from console.time(), you can use this line instead

console.log(`${lap} in: %c${(}ms`, 'color:blue');

// Usage: 
timer()              // set the start
// do something 
timer('built')       // logs 'built in: 591.815ms'
// do something
timer('copied')      // logs 'copied in: 0.065ms'
// do something
timer('compared')    // logs 'compared in: 36.41ms'

In my case, I perfer to use @ grammar suger and compile it with babel.
The problem of this method is that function has to be inside object.

Sample JS Code

function timer() {
    return (target, propertyKey, descriptor) => {
        const start =;
        let oldFunc = descriptor.value;

        descriptor.value = async function (){
            var result = await oldFunc.apply(this, arguments);
            console.log( - start);
            return result;

// Util function 
function delay(timeout) {
    return new Promise((resolve) => setTimeout(() => {
    }, timeout));

class Test {
    async test(timout) {
        await delay(timout)
        console.log("delay 1");
        await delay(timout)
        console.log("delay 2");

const t = new Test();

.babelrc (for babel 6)

    "plugins": [

The best way would be to use the performance hooks module. Although unstable, you can mark specific areas of your code and measure the duration between marked areas.

const { performance, PerformanceObserver } = require('perf_hooks');

const measures = []

const obs = new PerformanceObserver(list => measures.push(...list.getEntries()));
obs.observe({ entryTypes: ['measure'] });
const getEntriesByType = cb => cb(measures);

const doSomething = val => {
  performance.mark('beginning of the process');

  val *= 2;

  performance.mark('after multiplication');

  performance.measure('time taken', 'beginning of the process', 'after multiplication');

  getEntriesByType(entries => {
    entries.forEach(entry => console.log(entry));

  return val;


Try here

With performance

NodeJs: It is required to import the performance class

var time0 =; // Store the time at this point into time0

yourFunction();   // The function you're measuring time for 

var time1 =; // Store the time at this point into time1

console.log("youFunction took " + (time1 - time0) + " milliseconds to execute");

Using console.time


someFunction(); // Whatever is timed goes between the two "console.time"


  1. To start the timer use console.time("myTimer");
  2. Optional: To print the elapsed time, use
  3. Finally, to stop the timer and print the final
    time: console.timeEnd("myTimer");

You can read more about this on MDN and in the Node.js documentation.

Available on Chrome, Firefox, Opera and NodeJS. (not on Edge or Internet Explorer).

Use console.time('some label here') before the function and console.timeEnd('some label here') after the function. It will give you the running time of the function.

You should also consider reading about bigO notation. It may be better than timing functions for understanding what is going on

As previously stated check for and use built in timer. But if you want or need to write your own here is my two cents:

 * JavaScript Timer Object
 *      var now=timer['elapsed'](); 
 *      timer['stop']();
 *      timer['start']();
 *      timer['reset']();
 * @expose
 * @method timer
 * @return {number}
        /** @expose */
        reset:function(){return a=0}

//=-=|Google Advanced Optimized|=-=//
timer=function(){var;b=0;return{a:function(){return},start:function(){return},stop:function(){return},reset:function(){return a=0}}}();

Compilation was a success!

  • Original Size: 219 bytes gzipped (405 bytes uncompressed)
  • Compiled Size: 109 bytes gzipped (187 bytes uncompressed)
  • Saved 50.23% off the gzipped size (53.83% without gzip

The accepted answer is wrong !

Since JavaScript is asynchronous, the values of the variable end of the accepted answer would be wrong.

var start = new Date().getTime();

for (i = 0; i < 50000; ++i) {
// JavaScript is not waiting until the for is finished !!

var end = new Date().getTime();
var time = end - start;
alert('Execution time: ' + time); 

The execution of the for may be very fast so you can not see that the result is wrong. You can test it with a code doing some request :

var start = new Date().getTime();

for (i = 0; i < 50000; ++i) {
    url: '',
    success: function(){

var end = new Date().getTime();
var time = end - start;
alert('Execution time: ' + time); 

So the alert will prompt very quickly but in the console you’ll see that the ajax requests are continuing.

Here is how you should do it :

Leave a Comment